[ipxe-devel] trouble booting CentOS from Netapp iSCSI LUN

Scott Classen sclassen at lbl.gov
Thu Mar 3 05:12:45 UTC 2011

Please see comments below.

On Mar 2, 2011, at 5:58 PM, Michael Brown wrote:

> On Thursday 03 Mar 2011 00:21:58 Scott Classen wrote:
>> making a bit more progress. I think the tip about checking ACLs was
>> correct. I guess teh equivalent thing on a netapp are teh iniator
>> groups and mapping them to specific LUNs.
>> I had to add  iqn.2000-09.org.etherboot:unknown (presumably this is
>> how an iSCSI server sees iPXE?) to this initiator group on the netapp
> That is the current default, if no explicit initiator IQN is specified via 
> either DHCP option 203 or the "initiator-iqn" setting.

What is DHCP option 203? Can I add another line to my dhcpd.conf file to specify the iqn of my client?… like this:

host samos {
	hardware ethernet 00:25:90:99:99:99;
	if exists user-class and option user-class = "iPXE" {
                       filename "";
                       option root-path "iscsi:”;
			initiator-iqn “iqn.2010-2011.com.mycustominitiator"
               } else {
                       filename "/undionly.kpxe";

> I will have to change that default, since I don't own the etherboot.org 
> namespace, and I can't trust the owner not to cause problems in future.  This 
> may cause pain for people who have already added the default IQN to ACLs.  
> I'll try to minimise the pain by finally making the default IQN include the 
> local MAC address, which will be helpful since some targets refuse to let 
> multiple initiators connect simultaneously using the same (default) initiator 
> IQN.
>> now I get:
>> Registered as SAN device 0x80
>> Booting from SAN device 0x80
>> but the client machine just hangs indefinately.
> That usually indicates that the target disk doesn't contain a bootable OS.  
> What is installed on this disk?

Well I’m pretty sure a fresh install of CentOS 5.5
I performed a GUI-based installation and in the advanced partitioning section connected to the netapp, loaded the iSCSI LUNs and installed the root filesystem onto the NetApp LUN. The installation proceeded as expected and then I was prompted to reboot.
Of course this is where I have run into problems, but I think I’m making some good progress. At least iPXE can successfully connect to the NetApp I just need to figure out why the boot process freezes.

I think I might go through the CentOS install again and double check that /boot is also being installed on the iSCSI LUN.


More information about the ipxe-devel mailing list